CD74 Expression and Its Therapeutic Potential in Thyroid Carcinoma

S-P Cheng et al. CD74 in thyroid cancer 22:2 179–190 Research

CD74 expression and its therapeutic potential in thyroid carcinoma

Shih-Ping Cheng1,3,4, Chien-Liang Liu1,3, Ming-Jen Chen1,4,5, Ming-Nan Chien2,3, Ching-Hsiang Leung2,3, Chi-Hsin Lin5, Yi-Chiung Hsu6 and Jie-Jen Lee1,3,4

1Department of Surgery, 2Division of Endocrinology and Metabolism, Department of Medicine, MacKay Memorial Hospital and Mackay Medical College, Taipei, Taiwan 3Mackay Junior College of Medicine, Nursing, and Management, No. 92, Sheng-Ching Road, Peitou, Correspondence Taipei 11260, Taiwan should be addressed 4Department of Pharmacology, Graduate Institute of Medical Sciences, Taipei Medical University, Taipei, Taiwan to J-J Lee 5Department of Medical Research, MacKay Memorial Hospital, Taipei, Taiwan Email 6Institute of Statistical Science, Academia Sinica, Taipei, Taiwan [email protected]

Abstract

CD74, the invariant chain of major histocompatibility complex class II, is also a receptor for Key Words macrophage migration inhibitory factor (MIF). CD74 and MIF have been associated with " CD74 tumor progression and metastasis in hematologic and solid tumors. In this study, we found " macrophage migration that 60 and 65% of papillary thyroid cancers were positive for CD74 and MIF inhibitory factor immunohistochemical staining respectively. Anaplastic thyroid cancer was negative for MIF, " thyroid cancer but mostly positive for CD74 expression. Normal thyroid tissue and follicular adenomas were " antibody negative for CD74 expression. CD74 expression in papillary thyroid cancer was associated with larger tumor size (PZ0.043), extrathyroidal invasion (PZ0.021), advanced TNM stage

Endocrine-Related Cancer (PZ0.006), and higher MACIS score (PZ0.026). No clinicopathological parameter was associated with MIF expression. Treatment with anti-CD74 antibody in thyroid cancer cells inhibited cell growth, colony formation, cell migration and invasion, and vascular endothelial growth factor secretion. In contrast, treatment with recombinant MIF induced an increase in cell invasion. Anti-CD74 treatment reduced AKT phosphorylation and stimulated AMPK activation. Our ﬁndings suggest that CD74 overexpression in thyroid

cancer is associated with advanced tumor stage and may serve as a therapeutic target. Endocrine-Related Cancer (2015) 22, 179–190

Introduction

The majority of follicular cell-derived thyroid cancers is treatment of radioactive iodine-refractory differentiated differentiated carcinomas and can be treated successfully thyroid cancer. Further fundamental research is needed to with a combination of surgery and radioactive iodine discover new therapeutic targets with better effectiveness therapy. For patients with advanced thyroid cancer, whose and lower toxicity. disease progresses despite standard therapy, treatment CD74, also known as the invariant chain of major options are limited and survival is poor (Haugen & histocompatibility complex (MHC) class II, functions as Sherman 2013). Targeted therapies have begun receiving an MHC-II chaperone and participates in the trafﬁcking of attention in recent years. Currently, the multikinase MHC-II molecules in antigen-presenting cells (Borghese & inhibitor sorafenib is the only targeted agent approved Clanchy 2011). It has been shown that CD74 might be by the US Food and Drug Administration for the expressed independently of MHC-II, and CD74 can bind

http://erc.endocrinology-journals.org q 2015 Society for Endocrinology Published by Bioscientiﬁca Ltd. DOI: 10.1530/ERC-14-0269 Printed in Great Britain Downloaded from Bioscientifica.com at 09/29/2021 10:54:17AM via free access Research S-P Cheng et al. CD74 in thyroid cancer 22:2 180

to non-MHC-II proteins (Ogrinc et al. 1993, Henne et al. consisted of 103 papillary carcinomas and five anaplastic 1995). Moreover, CD74 is a cell membrane receptor carcinomas. Of the 103 papillary carcinomas, ten showed for macrophage migration inhibitory factor (MIF), a a follicular structure and were diagnosed as follicular pleiotropic cytokine that has pro-inflammatory and variant of papillary carcinoma. Lymph node metastasis pro-tumorigenic properties (Leng et al.2003). This was present in 58 (56%) of 103 patients with papillary ligand–receptor interaction may be an important link thyroid cancer. Among these patients, metastatic lymph between chronic inflammation and carcinogenesis (Bucala nodes from 27 (47%) patients were included for the & Donnelly 2007). The expression of MIF and/or CD74 has immunohistochemical study. In addition, sections of five been explored in several forms of cancer (Ren et al. 2005, follicular adenomas and five samples from Graves’ disease Xu et al. 2008, Nagata et al. 2009, Cheng et al. 2011a). In were submitted to immunohistochemical analysis. addition, MIF has been identified as a hypoxia-induced For western blot analysis, total proteins were extracted gene, and its expression serves to activate a proangiogenic from eight paired samples of papillary thyroid cancer, four transcriptional program (Winner et al. 2007). paired samples of follicular adenoma, four samples from In autoimmune thyroid disease, MIF production in Hashimoto’s thyroiditis, and four from Graves’ disease. response to thyroid antigens, but not to control antigens, All were collected during surgery and immediately has been described (Lamki et al. 1973). Recently, using snap-frozen in liquid nitrogen. a nationwide population-based study, we reported an increased risk for the development of thyroid cancer after Cell culture and reagents a diagnosis of thyroiditis (Liu et al. 2014). Considering that MIF upregulation is associated with autoimmune thyroid Human papillary thyroid carcinoma cell lines BCPAP and disease and certain types of cancer, it is therefore K1 were purchased from the German Collection of interesting to investigate whether a differential expression Microorganisms and Cell Cultures (DSMZ, Braunschweig, of MIF and CD74 exists in thyroid cancer. Furthermore, Germany) and the European Collection of Cell Cultures the restricted expression of CD74 among normal tissues (ECACC, Salisbury, UK) respectively. Both have been and its rapid internalization make CD74 an attractive authenticated to be unique thyroid cancer cell lines therapeutic target for cancer therapy (Hansen et al. 1996). (Schweppe et al. 2008). The expression of thyroid-specific Milatuzumab, a humanized MAB that recognizes CD74, is genes is shown in Supplementary Figure S1, see section currently under investigation. In a phase 1 trial, mono- on supplementary data given at the end of this article. Endocrine-Related Cancer therapy with milatuzumab was able to stabilize advanced BCPAP harbored the homozygous mutation of BRAF, multiple myeloma (Kaufman et al. 2013). In this study, we whereas K1 had the heterozygous mutation. BCPAP cells evaluated the significance of MIF and CD74 expressions in were cultured at 37 8C in a humidified atmosphere of 5%

thyroid cancer and the effects of anti-CD74 antibody on CO2 in RPMI 1640 medium (Gibco; Life Technologies) thyroid cancer cells. supplemented with 10% fetal bovine serum (FBS). K1 cells were cultured in DMEM (Gibco) mixed with Ham’s F12 (Sigma) and MCDB 105 (Sigma) medium in 2:1:1 Materials and methods proportions, supplemented with 10% FBS and 2 mM L-glutamine. Recombinant human MIF (#289-MF) was Patients and tissue samples purchased from R&D Systems (Minneapolis, MN, USA). This study was approved by the Institutional Review Board Anti-human CD74 antibody (specific for the extracellular (IRB) of MacKay Memorial Hospital (10MMHIS077). From domain of CD74, C-16; #sc-5438) was obtained from Santa January 2001 to November 2007, 334 patients underwent Cruz Biotechnology. thyroidectomy with therapeutic intent for newly diagnosed differentiated thyroid cancer at our institution. Cell viability and colony formation assay Incidental microcarcinomas were excluded. Eligible patients were de-identified and randomly selected for Cell growth was determined using the thiazolyl blue entry into the study. Each patient had complete clinical tetrazolium (MTT) bromide colorimetric assay. Briefly, and histopathological profile at diagnosis. Nonetheless, BCPAP and K1 cells (4000 cells/well) were seeded onto follow-up data were not available because of stringent 96-well plates and were allowed to grow in complete de-identification of patient records. Finally, a total of culture media for 24 h. Thereafter, the media was replaced 108 patients comprised our main study cohort. Lesions with culture media containing 1% FBS for a further

24–72 h. The cells were treated with isotype control, (Sysmex, Kobe, Japan), and counted under a light recombinant human MIF (200 ng/ml), or anti-CD74 microscope in five nonoverlapping fields. antibody (5 mg/ml). At the indicated time points, 100 ml (5 mg/ml) of MTT reagent (Sigma) was added to the cell Western blotting analysis culture and cells were incubated at 37 8C for 4 h. Formazan crystals converted from tetrazolium salts by viable cells Tissue or cellular proteins were extracted, quantified, and were solubilized with acidified isopropanol. The subjected to gel electrophoresis according to the standard absorbance was measured at 570 nm by a Varioskan procedures as we described previously (Cheng et al. 2012). Flash microplate reader (Thermo Fisher Scientific, A total of 30–50 mg of total protein was separated by Waltham, MA, USA). The absorbance of control cells was electrophoresis on 10–12% SDS polyacrylamide gels. defined as 100%. Fractionated proteins were transferred to a nitrocellulose Colony formation assay was performed as previously membrane, and transfer was controlled by Ponceau described (Chang et al. 2014). In brief, BCPAP (1000 cells/ staining. After transfer, the membrane was blocked with well) and K1 cells (500 cells/well) were seeded into six-well 5% skimmed milk or 5% BSA for 30 min at room plates, allowed to adhere for 24 h, and treated with isotype temperature. The proteins were probed with antibodies control, recombinant human MIF (200 ng/ml) or anti- against MIF (#ab55445; Abcam, Cambridge, MA, USA), CD74 antibody (5 mg/ml) from day 2. After 10–12 days, CD74 (#ab64103; Abcam), phospho-AKT (Thr308) (#4056; colonies were stained with 3% crystal violet and examined Cell Signaling, Danvers, MA, USA), phospho-AKT (Ser473) by microscopy. The colonies containing O50 cells were (#9271; Cell Signaling), total AKT (#2920; Cell Signaling), counted in five nonoverlapping fields. phospho-AMPKa (Thr172) (#2531; Cell Signaling), total AMPK (#2532; Cell Signaling), and b-actin (Sigma) at 4 8C overnight. The results were visualized by chemilumine- Wound healing assay scence with the Amersham ECL Detection System (GE Cell migration was determined using the wound healing Healthcare, Piscataway, NJ, USA). The blot signals were assay, which was modified from our published method quantified by densitometry (ImageJ Software) and (Cheng et al. 2011b). BCPAP and K1 cells were trypsinized, normalized to b-actin. counted, and seeded into six-well plates in culture media. After 24 h, monolayers of cells were wounded using a Endocrine-Related Cancer Cytokine detection sterile 300 ml micropipette tip to scratch on the confluent cells. The culture media were replaced with the medium Vascular endothelial growth factor (VEGF) concentrations (1% FBS) containing isotype control, recombinant human were determined using an ELISA (#DVE00; R&D Systems) MIF (200 ng/ml), or anti-CD74 antibody (5 mg/ml) according to the manufacturer’s instructions. The cell following scratch. The cells were photographed at 0, 20, culture supernatants were collected after each experiment 30, 40, and 48 h. Wound closure measurement was and centrifuged to remove cells and debris. The optical performed in four different fields using ImageJ Software density (OD) was measured at 450 nm and concentrations (National Institutes of Health, Bethesda, MD, USA). extrapolated from a standard curve. The sensitivity of the assays was 9 pg/ml.

Invasion assay Immunohistochemistry Cell invasion was assessed using the BioCoat Matrigel invasion chamber with 8 mm pore diameter (#354481; The expression of MIF and CD74 in tissue sections was Corning Life Sciences, Tewksbury, MA, USA). A total of studied by immunohistochemical staining as described 1!105 cells were seeded onto the inserts in culture media previously (Cheng et al. 2013). Briefly, 5 mm sections of containing 1% FBS with isotype control, recombinant paraffin-embedded tissue blocks were deparaffinized in human MIF (200 ng/ml), or anti-CD74 antibody (5 mg/ml). xylene and rehydrated in a graded series of ethanol. For The inserts were placed into a six-well plate with 10% antigen retrieval, the slides were immersed in a Tris–EDTA serum-containing culture medium as a chemoattractant. Buffer (10 mM Tris base, 1 mM EDTA solution, 0.05% The plates were incubated for 24–48 h at 37 8C. The cells Tween 20, and pH 9) and boiled for 20 min in a pressure that invaded the Matrigel matrix to the lower surface of cooker while maintaining the pressure. Endogenous the membrane were fixed and stained with Diff-Quik peroxidase activity was quenched with 3% hydrogen

peroxidase in methanol for 10 min. The slides were then Mann–Whitney U test or Student’s t-test. The Kendall’s t-b incubated with primary antibodies against MIF (#ab55445; test was used to determine the relationship between Abcam) and CD74 (#ab64103; Abcam) in humid chambers expressions of markers. Throughout the analysis, P values at 4 8C overnight. The slides were washed and further !0.05 (two-sided hypotheses) were considered to be incubated with goat-anti-mouse MACH2 HRP polymer statistically significant. Statistical analyses were performed (Biocare Medical, Concord, CA, USA) for 30 min at room using STATA for Windows, release 11 (Stata Corp., College temperature. The chromogenic reaction was done with Station, TX, USA). 0.02% 3,30-diaminobenzidine (Sigma) for 30 min. All slides were counterstained with hematoxylin, dehydrated, cleared, and cover slipped. Negative control slides in the Results absence of primary or secondary antibody were included Expression analysis using microarray database for each staining. Relative expression of MIF and CD74 was examined in different sets of thyroid carcinoma using microarray Evaluation of immunostaining database GSE3678, GSE3467, and GSE33630 downloaded Cytoplasmic immunoreactivity was evaluated semiquan- from the Gene Expression Omnibus database. There was titatively as previously reported (Cheng et al.2013). no significant difference in the expression level of MIF Staining intensity was scored as 0 (no staining), 1 (weak), between normal thyroid and papillary carcinoma, whereas 2 (moderate), and 3 (strong). Staining proportion was CD74 expression was significantly increased in papillary scored as 0 (!1%), 1 (1–25%), 2 (26–50%), and 3 (O50%), carcinoma compared with normal thyroid (Fig. 1). according to the percentage of positively stained cells. Anaplastic cancer samples showed significantly lower Multiplied scores of intensity and proportion were used as

the final staining score. Positive expression was defined by A NS * final staining scores 6 and 9, whereas the remaining cases 800 5000

(ﬁnal scores 0–4) were classiﬁed as negative expression. 4000 600 3000 400

expression 2000 expression Analysis of publicly available gene expression datasets 200 MIF

CD74 1000 Endocrine-Related Cancer We accessed the National Center for Biotechnology Infor- 0 0 Normal PTC Normal PTC mation Gene Expression Omnibus (http://www.ncbi.nlm. nih.gov/geo/) and identiﬁed publicly available thyroid B 14 NS 15 ** 14 cancer datasets. GSE3678 comprises gene expression data 13 13 of seven paired samples of papillary thyroid cancer. GSE3467 12

expression 12 (He et al.2005) comprises gene expression data of nine expression 11 MIF paired samples of papillary thyroid cancer. GSE33630 CD74 11 10 10 (Hebrant et al.2012) comprises gene expression data of Normal PTC Normal PTC 45 normal thyroid and 49 papillary thyroid cancer samples C ** ** obtained from Ukraine via the Chernobyl Tissue Bank, as NS 14 16 *** well as 11 anaplastic thyroid cancer samples from different 13 14 hospitals in France and Belgium. Affymetrix Human 12 12 expression

Genome U133 Plus 2.0 Arrays were used in these datasets. expression 11 10 MIF CD74

10 8 Statistical analyses Normal PTC ATC Normal PTC ATC

Categorical patient characteristics are expressed as Figure 1 proportions, while continuous variables are expressed as Gene expression levels of macrophage migration inhibitory factor (MIF) medians with 25th and 75th percentiles (deﬁning the and CD74 in normal thyroid, papillary thyroid cancer (PTC), and anaplastic interquartile range). c2, Fisher’s exact test, or Cochran– thyroid cancer (ATC). Three different datasets were examined: (A) GSE3678 dataset, (B) GSE3467 dataset (He et al. 2005), and (C) GSE33630 dataset Armitage trend test was used to compare categorical (Hebrant et al. 2012). P values were determined by the Mann–Whitney variables. Continuous variables were compared using the U test. *P!0.05, **P!0.01, and ***P!0.001; NS, non-signiﬁcant.

expression of MIF and higher expression of CD74 as In the five anaplastic cancers, three showed no MIF compared with normal thyroid specimens (Fig. 1C). staining within anaplastic components, and two had faint staining, which was also classified as negative expression. Positive CD74 immunoreactivity was present in 4 (80%) Tissue expression in thyroid cancer of five cases. Our findings are consistent with those of Immunohistochemical staining for MIF and CD74 was microarray gene expression data. Additional five follicular performed in 103 papillary cancers, 27 corresponding adenomas and five samples from Graves’ disease were metastatic lymph nodes, and five anaplastic cancers. Based analyzed. MIF immunoreactivity was positive in four on our predetermined criteria, 67 (65%) and 62 (60%) follicular adenomas and three samples from Graves’ papillary cancers were positive for MIF and CD74 staining disease respectively. No CD74 positive staining was respectively (Fig. 2A). In all, 44 (43%) exhibited both MIF noted in these tissue samples of benign thyroid lesions and CD74 expression. There was no correlation between the (Supplementary Figure S2, see section on supplementary expression of MIF and CD74 (PZ0.124). Among patients data given at the end of this article). Our findings suggest with papillary cancer, lymph nodes positive for metastasis that CD74 overexpression in thyroid cancer is relatively were randomly selected from 27 patients and subjected to specific to malignant transformation. immunohistochemical analysis. We observed that MIF Western blotting analysis was used to evaluate the immunoreactivity appeared in 25 (93%) of 27 patients, expression of MIF and CD74 in eight paired (normal and whereas positive CD74 expression was evident in 21 (78%). tumor) samples of papillary thyroid cancer (Fig. 2B). There There was no significant correlation in MIF or CD74 was a trend toward increased MIF and CD74 expression expression between the primary tumor and metastatic in the tumor part. It is noteworthy that CD74 expression lymph nodes (PZ0.828 and PZ0.946 respectively). was barely detectable in the normal thyroid tissue.

A Normal PTCLymph node ATC

MIF Endocrine-Related Cancer

CD74

NTNTNTNT NTNT NT 2 MIF 38 kDa

CD74 34 kDa 1

β Normalized T/N ratio -actin 42 kDa 0 MIF CD74

Figure 2 Tissue expression of macrophage migration inhibitory factor (MIF) and blotting analysis in paired PTC samples (T, tumor; N, adjacent normal CD74 in thyroid cancer. (A) Representative microphotographs of immuno- thyroid tissue). Bands were quantified by densitometry and normalized to histochemical staining in normal thyroid, papillary thyroid cancer (PTC), the b-actin levels. T:N ratios were plotted on a vertical scatter plot. metastatic lymph node, and anaplastic thyroid cancer (ATC). Original Horizontal bars indicate the mean for each group. A full colour version of magnification, 100! for normal thyroid, 200! for PTC, 40! for lymph this figure is available at http://dx.doi.org/10.1530/ERC-14-0269. node, and 100! for ATC. (B) Protein expression determined by western

Table 1 Correlation of macrophage migration inhibitory concurrent lymphocytic thyroiditis. Given that CD74 factor (MIF) expression with clinicopathological parameters in is overexpressed in thyroid cancer and associated with patients with papillary thyroid carcinoma (nZ103) advanced tumor stage, we hypothesize that CD74 might be a potential therapeutic target in thyroid cancer. MIF (C)(nZ67) MIF (K)(nZ36) P value

Female 52 (78%) 29 (81%) 0.728 Age (years)a 41 (33–49) 40 (30–51) 0.561 Reduced cell growth with anti-CD74 antibody treatment Hashimoto’s thyroiditis 3 (4%) 3 (8%) 0.419 Tumor size (cm)a 2.5 (2.0–3.2) 2.1 (1.8–2.6) 0.111 Two papillary thyroid cancer cell lines, BCPAP and K1, were b Extrathyroidal invasion 0.100 evaluated for the expression of CD74. Both thyroid cancer No 41 (61%) 27 (75%) Microscopic 19 (28%) 9 (25%) cells were positive for CD74 expression (Supplementary Macroscopic 7 (10%) 0 (0%) Figure S3). Treatment with MIF did not alter the CD74 Multifocality 22 (33%) 15 (42%) 0.373 expression, whereas anti-CD74 neutralizing antibody Lymph node metastasisb 0.610 N0 29 (43%) 16 (44%) (which binds the CD74 extracellular domain) remarkably N1a 25 (37%) 16 (44%) decreased the CD74 expression in both cell lines. N1b 13 (19%) 4 (11%) We then tested the effects of MIF and anti-CD74 Distant metastasis 2 (3%) 0 (0%) 0.541 TNM stageb 0.189 antibody on cell growth. Both cells were treated with Stage 1 46 (69%) 29 (81%) vehicle (PBS) control, isotype control, recombinant human Stage 2 5 (7%) 2 (6%) MIF, or anti-CD74 antibody. The results of vehicle control Stage 3 9 (13%) 3 (8%) Stage 4 7 (10%) 2 (6%) and isotype control were similar, and thus only the results a MACIS score 4.8 (4.0–5.9) 4.5 (3.9–5.5) 0.252 of isotype control were shown. As shown in Fig. 3A, MACIS R6 15 (22%) 4 (11%) 0.191 treatment with anti-CD74 antibody signiﬁcantly

Data are expressed as number (percentage) or median (interquartile range). suppressed cell growth after 48 h. There was no signiﬁcant MACIS, Mayo Clinic’s metastasis, patient age, completeness of resection, local invasion, and tumor size. aMann–Whitney U test. bCochran–Armitage trend test. Table 2 Correlation of CD74 expression with clinicopathological parameters in patients with papillary thyroid carcinoma (nZ103) An inconsistent pattern of MIF expression was observed Endocrine-Related Cancer in samples of follicular adenoma, Hashimoto’s thyroiditis, CD74 (C) CD74 (K) and Graves’ disease (Supplementary Figure S2). These (nZ62) (nZ41) P value benign thyroid lesions did not show CD74 expression. Female 50 (81%) 31 (76%) 0.542 Age (years)a 42 (33–51) 37 (31–47) 0.103 Hashimoto’s thyroiditis 4 (6%) 2 (5%) 1.000 a Clinicopathological correlations Tumor size (cm) 2.5 (2.0–3.2) 2.1 (1.5–2.5) 0.043 Extrathyroidal invasionb 0.021 To investigate the potential role of MIF and CD74 in No 36 (58%) 32 (78%) Microscopic 19 (31%) 9 (22%) thyroid cancer, the 103 samples of papillary cancer were Macroscopic 7 (11%) 0 (0%) divided into two groups with positive and negative Multifocality 20 (32%) 17 (41%) 0.341 b MIF/CD74 expression and compared for a series of Lymph node metastasis 0.930 N0 28 (45%) 17 (41%) clinicopathological parameters. As given in Table 1,we N1a 22 (35%) 19 (46%) have not observed any parameter that was associated with N1b 12 (19%) 5 (12%) the MIF expression. However, papillary cancer positive for Distant metastasis 2 (3%) 0 (0%) 0.516 TNM stageb 0.006 CD74 expression was associated with larger tumor size, Stage 1 39 (63%) 36 (88%) higher rate of extrathyroidal invasion, and more advanced Stage 2 6 (10%) 1 (2%) Stage 3 9 (15%) 3 (7%) TNM stage (Table 2). MACIS scoring system developed at Stage 4 8 (13%) 1 (2%) the Mayo Clinic uses metastasis, age, completeness of MACIS scorea 5.1 (4.0–6.0) 4.4 (3.9–5.2) 0.026 resection, local invasion, and tumor size to stratify patients MACIS R6 16 (26%) 3 (7%) 0.020 by different survival (Hay et al. 1993). In accordance, CD74- Data are expressed as number (percentage) or median (interquartile range). positive thyroid cancer had signiﬁcantly higher MACIS MACIS, Mayo Clinic’s metastasis, patient age, completeness of resection, scores (PZ0.026). The expression of MIF or CD74 was not local invasion, and tumor size. aMann-Whitney test. associated with follicular variant or the presence of bCochran–Armitage trend test.

A BCPAP K1 Control Control 150 MIF 150 MIF CD74Ab CD74Ab

100 100 ** *** ** *** 50 50 Cell growth (%) Cell growth Cell growth (%) Cell growth

0 0 24 h 48 h 72 h 24 h 48 h 72 h

B BCPAP K1

80 200

60 150 * ** 40 100 Colonies Colonies 20 *** 50

0 0 Control MIF CD74Ab Control MIF CD74Ab

Figure 3 Effects of CD74 modulation on cell growth and colony formation in human was calculated by comparing results with control cells (100% viable). Endocrine-Related Cancer thyroid cancer cells. (A) BCPAP and K1 cells were treated with isotype (B) Quantiﬁcation of colonies formed following treatment with isotype control, recombinant human macrophage migration inhibitory factor control, MIF, or CD74Ab in BCPAP and K1 cells. P values were determined by (MIF, 200 ng/ml), or anti-CD74 antibody (CD74Ab, 5 mg/ml) for 24–72 h. the Student’s t-test. *P!0.05, **P!0.01, and ***P!0.001. A full colour Cell viability, determined with thiazolyl blue tetrazolium (MTT) assay, version of this ﬁgure is available at http://dx.doi.org/10.1530/ERC-14-0269.

effect on cell growth with MIF treatment. In concordance cell migration. Similarly, treatment with anti-CD74 with the results of MTT assay, treatment with anti-CD74 antibody signiﬁcantly reduced the invasiveness through antibody inhibited anchorage-dependent colony formation matrigel (BCPAP, 48% and K1, 35%) compared with by 86% in BCPAP cells and 35% in K1 cells (Fig. 3B). In controls (Fig. 4B). Interestingly, treatment with recombi- addition, the size of colonies in anti-CD74 antibody-treated nant human MIF led to a remarkable increase in invasion cells was smaller as compared with isotype control-treated in both BCPAP (67% increase) and K1 (69% increase) cells. cells. Paradoxically, clonogenic assays showed that MIF treatment signiﬁcantly reduced the colony numbers by 32% Decreased VEGF secretion with anti-CD74 antibody in BCPAP cells but not in K1 cells. treatment

Previous study showed that MIF induced VEGF secretion Decreased cell migration and invasion with in a dose-dependent manner in cervical cancer cells, anti-CD74 antibody treatment which could be inhibited by anti-CD74 antibody In the same experimental conditions using MTT assay, cell (Cheng et al.2011a). To study the effects on VEGF migration and invasion were evaluated in thyroid cancer secretion in thyroid cancer, BCPAP and K1 cells were cells. With treatment with anti-CD74 antibody, cell treated with the same experimental protocol previously migration was reduced by 34% in BCPAP cells and 29% used. As shown in Fig. 5A, MIF did not stimulate VEGF in K1 cells (Fig. 4A). MIF treatment had no effect on secretion in thyroid cancer cells. However, treatment with

A BCPAP K1

0 h 0 h

20 h 20 h

60 60

40 * 40 **

% closure 20 % closure 20

0 0 Control MIF CD74Ab Control MIF CD74Ab

B BCPAP K1

300 ** 400 ** 300 Endocrine-Related Cancer 200 200 ** 100 **

Invaded cells Invaded cells Invaded 100

0 0 Control MIF CD74Ab Control MIF CD74Ab

Figure 4 Effects of CD74 modulation on cellular migration and invasion in human (CD74Ab, 5 mg/ml). (B) Quantification of invaded cells following treatment thyroid cancer cells. (A) Quantification of wound closure in BCPAP and with isotype control, MIF, or CD74Ab for 48 h. P values were determined by K1 cells treated with isotype control, recombinant human macrophage the Student’s t-test. *P!0.05 and **P!0.01. A full colour version of this migration inhibitory factor (MIF, 200 ng/ml), or anti-CD74 antibody figure is available at http://dx.doi.org/10.1530/ERC-14-0269.

anti-CD74 antibody modestly decreased VEGF secretion proliferation (Starlets et al. 2006). Furthermore, impaired by 29% in BCPAP cells and 12% in K1 cells. AMPK activation was observed in the hearts of CD74- knockout mice during ischemia (Qi et al. 2014). Decreased AKT Thr308 phosphorylation was observed following anti- Survival pathways involved in CD74 modulation CD74 antibody treatment, while AKT Ser473 phosphoryl- To investigate the mechanisms involved in CD74 modu- ation was slightly reduced (Fig. 5B). In addition, AMPK lation, the expression of the phosphorylation of AKT and phosphorylation increased upon stimulation with anti- AMPK was determined following treatment with anti- CD74 antibody in both cell lines. There was no change in CD74 antibody. Stimulation of surface CD74 induces a the expression of total AKT or total AMPK. These results signaling cascade resulting in AKT activation and cell indicate that PI3K/AKT and AMPK pathways may play a

A BCPAP K1 Control 3000 Control MIF MIF 600 CD74Ab CD74Ab * 2000 400 * 1000 VEGF (pg/ml) 200 VEGF (pg/ml) *

0 0 0 h 24 h 48 h 0 h 24 h 48 h

B BCPAP K1

CD74Ab 0 15 30 60 120 0 15 30 60 120 (min)

pAKT (Thr308) 60 kDa

pAKT (Ser473) 60 kDa

Total AKT 60 kDa

pAMPK (Thr172) 62 kDa

Total AMPK 62 kDa

β-actin 42 kDa

Figure 5 Effects of CD74 modulation on vascular endothelial growth factor (VEGF) medium was analyzed by ELISA. P values were determined by the Student’s secretion and various kinases in human thyroid cancer cells. (A) BCPAP and t-test. *P!0.05. (B) Cells were treated with CD74Ab for various times. K1 cells were treated with isotype control, recombinant human macro- Protein extraction and western blotting were performed as described in Endocrine-Related Cancer phage migration inhibitory factor (MIF, 200 ng/ml), or anti-CD74 antibody Materials and methods section. (CD74Ab, 5 mg/ml) for 24–48 h. The secretion of VEGF in the culture

role in the inhibitory effects seen following treatment with different implications from exogenous MIF stimulation. anti-CD74 antibody. For instance, breast cancer with abundant cytosolic MIF expression was associated with better disease-specific overall and recurrence-free survival (Verjans et al. 2009). Discussion Considering that there was no significant difference In this study, we for the first time demonstrated that CD74 in clinicopathological parameters between MIF-positive was overexpressed in tumorous parts of thyroid malig- and MIF-negative tumors, MIF probably plays a minor role nancy and was associated with advanced tumor stage. Our in thyroid cancer development and progression. CD74 initial hypothesis was that MIF may represent a link instead might be more appropriate for therapeutic between inflammation and thyroid cancer. MIF binds to intervention in thyroid cancer. CD74, and correlations between the expression of CD74 We observed that treatment with anti-CD74 antibody and MIF have been reported (Cheng et al. 2011a, Zheng inhibited cell growth, migration, and invasion in thyroid et al.2012). However, based on our results, CD74 cancer cells. In DU-145 prostate cancer cells, anti-CD74 expression did not correlate with MIF expression in treatment reduced cell viability to the extent similar to thyroid cancer. Furthermore, anaplastic thyroid cancer that observed following knockdown of CD74 by RNA showed a decreased MIF expression. CD74 expression is interference (Meyer-Siegler et al. 2006). Furthermore, anti- generally observed in tumor with advanced stage and CD74 treatment successfully suppressed cell invasion in worse survival (Nagata et al. 2009, Zheng et al. 2012). On prostate cancer cells expressing cell surface CD74. In MKN- the other hand, intracellular MIF expression may have 45 gastric cancer cells, lipopolysaccharide-induced cell

proliferation was similarly reduced with either knock- et al. 2006). CD74 (but not MIF) stimulation induced down of CD74 or anti-CD74 treatment (Zheng et al. 2012). upregulation of monocyte chemoattractant protein 1, Antibody therapeutics in cancer is used to specifically which was prevented by an AKT inhibitor (Martin-Ventura target a molecule with antigenic differences between et al. 2009). Our data are in concordance with these results normal and malignant tissues (Sliwkowski & Mellman and suggested an association between inhibition of 2013). Moreover, CD74 is a suitable carrier for antibody– PI3K/AKT pathway and suppressive effects following drug conjugates for its high expression on neoplastic cells treatment with anti-CD74 antibody. and rapid internalization. Various chemotherapeutic or AMPK is a master energy sensor and functions to radioactive agents have been used to prepare a conjugate monitor and maintain cellular and organismal meta- in hematologic and solid tumors (Govindan et al. 2013). bolism. AKT negatively regulates AMPK by decreasing Despite our intriguing findings suggesting relatively the AMP:ATP ratio (Hahn-Windgassen et al. 2005). In this specific expression of CD74 in malignant thyroid cancer respect, we found that anti-CD74 treatment inhibited AKT tissues, further in vivo studies are warranted to evaluate the phosphorylation and meanwhile activated the AMPK efficacy and safety of targeting CD74 in thyroid cancer. pathway. MIF stimulation activates the cardioprotective Cancer-induced immune stimulation and suppression AMPK pathway during ischemia (Miller et al.2008). may lead to local and systemic cytokine alterations. Nonetheless, AMPK activation had a greater degree of K K K K Among cytokine alterations associated with malignancy, impairment in CD74 / than MIF / hearts during significantly increased serum MIF concentrations have ischemia (Qi et al. 2014). The sequence or interaction of been reported in several types of cancer (Lippitz 2013). AKT inhibition and AMPK activation could not be In esophageal and breast cancer, tissue MIF expression determined in this study. In the opposite direction, correlated with microvessel density (Ren et al. 2005, Xu AMPK may suppress AKT activity through activation of et al.2008). In cervical cancer, either MIF or CD74 PTEN or protein phosphatase 2A (Huang et al. 2008, Kim expression was positively associated with higher micro- et al. 2009). AMPK agonists are currently under investi- vessel density (Cheng et al. 2011a). MIF has been shown gation for cancer treatment. Metformin use in diabetic to interact with and stabilize hypoxia-inducible factor 1 patients with thyroid cancer is associated with higher alpha, a master regulator of hypoxic/ischemic vascular remission rate (Klubo-Gwiezdzinska et al. 2013). The role responses (Winner et al.2007). In this study, MIF of AMPK activation in tumor suppressive effects associated treatment did not induce VEGF secretion in thyroid Endocrine-Related Cancer with CD74 modulation remains to be clarified. cancer cells. In contrast, anti-CD74 antibody modestly One limitation of our study is that follow-up data reduced the secretion of VEGF. Our findings are consistent are not available for analysis. Therefore, we could not with those of Kindt et al. (2014) that CD74 knockdown affirm whether patients with a CD74-positive tumor have reduced VEGF production from SCCVII cells. A large body a worse overall or disease-specific survival. Furthermore, of evidence indicates that VEGF plays a critical role in angiogenesis and metastasis (Phay & Ringel 2013). Clinical the number of benign thyroid lesions studied is small. applications of VEGF in thyroid cancer have included The specificity of CD74 overexpression in malignant diagnosis, prediction of prognosis, and treatment (Lin & thyroid tissue needs to be verified in a larger series of Chao 2005). Collectively, these findings provide further tumors. We observed CD74 overexpression in 60% of the evidence that CD74 may serve as a potential target for the primary papillary cancer and 78% of the corresponding intervention of thyroid cancer. metastatic lymph nodes. There was no association of AKT signaling plays an important role in the pro- CD74 expression between the primary tumor and meta- gression of thyroid cancer (Shinohara et al. 2007). It was static nodes. It is unclear whether CD74 overexpression shown that MIF prevents apoptosis and promotes tumor develops de novo during metastasis or CD74-positive cells cell survival by directly activating the AKT pathway (Lue carrying greater migratory/invading capacity are selected et al. 2007). The function of AKT is regulated by in the metastatic process. phosphorylation on two sites. Maximal AKT activity In summary, we report the characterization of CD74 requires Thr308 phosphorylation mediated by PDK1 and expression in thyroid cancer and demonstrate that Ser473 phosphorylation by mTOR complex 2 (mTORC2; treatment with anti-CD74 antibody effectively modulates Manning & Cantley 2007). It has been shown that malignant cell phenotype. Although the findings appear stimulation of surface CD74 induced a signaling cascade to rebut the initial hypothesis that MIF/CD74 mediates resulting in AKT activation and cell proliferation (Starlets the link between inflammation and thyroid cancer, our

results indicate that CD74 may serve as a therapeutic cellular ATP level and AMPK activity. Journal of Biological Chemistry 280 target molecule in advanced thyroid cancer. 32081–32089. (doi:10.1074/jbc.M502876200) Hansen HJ, Ong GL, Diril H, Valdez A, Roche PA, Griffiths GL, Goldenberg DM & Mattes MJ 1996 Internalization and catabolism of radiolabelled antibodies to the MHC class-II invariant chain by B-cell lymphomas. Biochemical Journal 320 293–300. Supplementary data Haugen BR & Sherman SI 2013 Evolving approaches to patients with This is linked to the online version of the paper at http://dx.doi.org/10.1530/ advanced differentiated thyroid cancer. Endocrine Reviews 34 439–455. ERC-14-0269. (doi:10.1210/er.2012-1038) Hay ID, Bergstralh EJ, Goellner JR, Ebersold JR & Grant CS 1993 Predicting outcome in papillary thyroid carcinoma: development of Declaration of interest a reliable prognostic scoring system in a cohort of 1779 patients The authors declare that there is no conflict of interest that could be surgically treated at one institution during 1940 through 1989. Surgery perceived as prejudicing the impartiality of the research reported. 114 1050–1057. He H, Jazdzewski K, Li W, Liyanarachchi S, Nagy R, Volinia S, Calin GA, Liu CG, Franssila K, Suster S et al. 2005 The role of microRNA genes in papillary thyroid carcinoma. PNAS 102 19075–19080. (doi:10.1073/ Funding pnas.0509603102) This study was supported by grants from the National Science Council of Hebrant A, Dom G, Dewaele M, Andry G, Tresallet C, Leteurtre E, Dumont JE Taiwan (NSC-100-2314-B-195-001-MY3 and MOST-103-2314-B-195-015- & Maenhaut C 2012 mRNA expression in papillary and anaplastic MY3) and MacKay Memorial Hospital (MMH-10303 and MMH-10423). thyroid carcinoma: molecular anatomy of a killing switch. PLoS ONE 7 e37807. (doi:10.1371/journal.pone.0037807) Henne C, Schwenk F, Koch N & Moller P 1995 Surface expression of the invariant chain (CD74) is independent of concomitant expression of Acknowledgements major histocompatibility complex class II antigens. Immunology 84 Part of the data of this study was presented in oral form at the 49th 177–182. International Congress of European Society for Surgical Research, Huang X, Wullschleger S, Shapiro N, McGuire VA, Sakamoto K, Woods YL, Budapest, Hungary, May 2014. McBurnie W, Fleming S & Alessi DR 2008 Important role of the LKB1–AMPK pathway in suppressing tumorigenesis in PTEN-deficient mice. Biochemical Journal 412 211–221. (doi:10.1042/BJ20080557) Kaufman JL, Niesvizky R, Stadtmauer EA, Chanan-Khan A, Siegel D, Horne H, References Wegener WA & Goldenberg DM 2013 Phase I, multicentre, dose-escalation trial of monotherapy with milatuzumab (humanized Borghese F & Clanchy FI 2011 CD74: an emerging opportunity as a anti-CD74 monoclonal antibody) in relapsed or refractory multiple therapeutic target in cancer and autoimmune disease. Expert Opinion myeloma. British Journal of Haematology 163 478–486. (doi:10.1111/ on Therapeutic Targets 15 237–251. (doi:10.1517/14728222.2011. bjh.12565) Kim KY, Baek A, Hwang JE, Choi YA, Jeong J, Lee MS, Cho DH, Lim JS, Kim KI

Endocrine-Related Cancer 550879) Bucala R & Donnelly SC 2007 Macrophage migration inhibitory factor: a & Yang Y 2009 Adiponectin-activated AMPK stimulates depho- probable link between inﬂammation and cancer. Immunity 26 281–285. sphorylation of AKT through protein phosphatase 2A activation. (doi:10.1016/j.immuni.2007.03.005) Cancer Research 69 4018–4026. (doi:10.1158/0008-5472.CAN-08-2641) Chang YC, Hsu YC, Liu CL, Huang SY, Hu MC & Cheng SP 2014 Local Kindt N, Lechien JR, Nonclercq D, Laurent G & Saussez S 2014 Involvement anesthetics induce apoptosis in human thyroid cancer cells through of CD74 in head and neck squamous cell carcinomas. Journal of Cancer the mitogen-activated protein kinase pathway. PLoS ONE 9 e89563. Research and Clinical Oncology 140 937–947. (doi:10.1007/s00432-014- (doi:10.1371/journal.pone.0089563) 1648-9) Cheng RJ, Deng WG, Niu CB, Li YY & Fu Y 2011a Expression of macrophage Klubo-Gwiezdzinska J, Costello J Jr, Patel A, Bauer A, Jensen K, Mete M, migration inhibitory factor and CD74 in cervical squamous cell Burman KD, Wartofsky L & Vasko V 2013 Treatment with metformin is carcinoma. International Journal of Gynecological Cancer 21 1004–1012. associated with higher remission rate in diabetic patients with thyroid (doi:10.1097/IGC.0b013e31821c45b7) cancer. Journal of Clinical Endocrinology and Metabolism 98 3269–3279. Cheng SP, Yin PH, Hsu YC, Chang YC, Huang SY, Lee JJ & Chi CW 2011b (doi:10.1210/jc.2012-3799) Leptin enhances migration of human papillary thyroid cancer cells Lamki L, Row VV & Volpe R 1973 Cell-mediated immunity in Graves’ through the PI3K/AKT and MEK/ERK signaling pathways. Oncology disease and in Hashimoto’s thyroiditis as shown by the demonstration Reports 26 1265–1271. (doi:10.3892/or.2011.1388) of migration inhibition factor (MIF). Journal of Clinical Endocrinology and Cheng SP, Liu CL, Hsu YC, Chang YC, Huang SY & Lee JJ 2012 Regulation of Metabolism 36 358–364. (doi:10.1210/jcem-36-2-358) leptin receptor expression in human papillary thyroid cancer cells. Leng L, Metz CN, Fang Y, Xu J, Donnelly S, Baugh J, Delohery T, Chen Y, Biomedicine & Pharmacotherapy 66 469–473. (doi:10.1016/j.biopha. Mitchell RA & Bucala R 2003 MIF signal transduction initiated by 2012.03.008) binding to CD74. Journal of Experimental Medicine 197 1467–1476. Cheng SP, Liu CL, Hsu YC, Chang YC, Huang SY & Lee JJ 2013 Expression (doi:10.1084/jem.20030286) and biologic signiﬁcance of adiponectin receptors in papillary thyroid Lin JD & Chao TC 2005 Vascular endothelial growth factor in thyroid carcinoma. Cell Biochemistry and Biophysics 65 203–210. (doi:10.1007/ cancers. Cancer Biotherapy & Radiopharmaceuticals 20 648–661. s12013-012-9419-1) (doi:10.1089/cbr.2005.20.648) Govindan SV, Cardillo TM, Sharkey RM, Tat F, Gold DV & Goldenberg DM Lippitz BE 2013 Cytokine patterns in patients with cancer: a systematic 2013 Milatuzumab–SN-38 conjugates for the treatment of CD74C review. Lancet. Oncology 14 e218–e228. (doi:10.1016/S1470- cancers. Molecular Cancer Therapeutics 12 968–978. (doi:10.1158/1535- 2045(12)70582-X) 7163.MCT-12-1170) Liu CL, Cheng SP, Lin HW & Lai YL 2014 Risk of thyroid cancer in patients Hahn-Windgassen A, Nogueira V, Chen CC, Skeen JE, Sonenberg N & Hay N with thyroiditis: a population-based cohort study. Annals of Surgical 2005 Akt activates the mammalian target of rapamycin by regulating Oncology 21 843–849. (doi:10.1245/s10434-013-3363-1)

Lue H, Thiele M, Franz J, Dahl E, Speckgens S, Leng L, Fingerle-Rowson G, Ren Y, Law S, Huang X, Lee PY, Bacher M, Srivastava G & Wong J 2005 Bucala R, Luscher B & Bernhagen J 2007 Macrophage migration Macrophage migration inhibitory factor stimulates angiogenic factor inhibitory factor (MIF) promotes cell survival by activation of the Akt expression and correlates with differentiation and lymph node status in pathway and role for CSN5/JAB1 in the control of autocrine MIF patients with esophageal squamous cell carcinoma. Annals of Surgery activity. Oncogene 26 5046–5059. (doi:10.1038/sj.onc.1210318) 242 55–63. (doi:10.1097/01.sla.0000168555.97710.bb) Manning BD & Cantley LC 2007 AKT/PKB signaling: navigating down- Schweppe RE, Klopper JP, Korch C, Pugazhenthi U, Benezra M, Knauf JA, stream. Cell 129 1261–1274. (doi:10.1016/j.cell.2007.06.009) Fagin JA, Marlow LA, Copland JA, Smallridge RC et al. 2008 Martin-Ventura JL, Madrigal-Matute J, Munoz-Garcia B, Blanco-Colio LM, Deoxyribonucleic acid profiling analysis of 40 human thyroid cancer Van Oostrom M, Zalba G, Fortuno A, Gomez-Guerrero C, Ortega L, cell lines reveals cross-contamination resulting in cell line redundancy Ortiz A et al. 2009 Increased CD74 expression in human atherosclerotic and misidentification. Journal of Clinical Endocrinology and Metabolism plaques: contribution to inflammatory responses in vascular cells. 93 4331–4341. (doi:10.1210/jc.2008-1102) Cardiovascular Research 83 586–594. (doi:10.1093/cvr/cvp141) Shinohara M, Chung YJ, Saji M & Ringel MD 2007 AKT in thyroid Meyer-Siegler KL, Iczkowski KA, Leng L, Bucala R & Vera PL 2006 Inhibition tumorigenesis and progression. Endocrinology 148 942–947. of macrophage migration inhibitory factor or its receptor (CD74) (doi:10.1210/en.2006-0937) attenuates growth and invasion of DU-145 prostate cancer cells. Journal of Sliwkowski MX & Mellman I 2013 Antibody therapeutics in cancer. Science Immunology 177 8730–8739. (doi:10.4049/jimmunol.177.12.8730) 341 1192–1198. (doi:10.1126/science.1241145) Miller EJ, Li J, Leng L, McDonald C, Atsumi T, Bucala R & Young LH 2008 Starlets D, Gore Y, Binsky I, Haran M, Harpaz N, Shvidel L, Becker-Herman S, Macrophage migration inhibitory factor stimulates AMP-activated Berrebi A & Shachar I 2006 Cell-surface CD74 initiates a signaling protein kinase in the ischaemic heart. Nature 451 578–582. cascade leading to cell proliferation and survival. Blood 107 4807–4816. (doi:10.1038/nature06504) (doi:10.1182/blood-2005-11-4334) Nagata S, Jin YF, Yoshizato K, Tomoeda M, Song M, Iizuka N, Kitamura M, Verjans E, Noetzel E, Bektas N, Schutz AK, Lue H, Lennartz B, Hartmann A, Takahashi H, Eguchi H, Ohigashi H et al. 2009 CD74 is a novel Dahl E & Bernhagen J 2009 Dual role of macrophage migration prognostic factor for patients with pancreatic cancer receiving multi- inhibitory factor (MIF) in human breast cancer. BMC Cancer 9 230. modal therapy. Annals of Surgical Oncology 16 2531–2538. (doi:10.1245/ (doi:10.1186/1471-2407-9-230) s10434-009-0532-3) Winner M, Koong AC, Rendon BE, Zundel W & Mitchell RA 2007 Ogrinc T, Dolenc I, Ritonja A & Turk V 1993 Purification of the complex of Amplification of tumor hypoxic responses by macrophage migration cathepsin L and the MHC class II-associated invariant chain fragment inhibitory factor-dependent hypoxia-inducible factor stabilization. from human kidney. FEBS letters 336 555–559. (doi:10.1016/0014- Cancer Research 67 186–193. (doi:10.1158/0008-5472.CAN-06-3292) 5793(93)80875-U) Xu X, Wang B, Ye C, Yao C, Lin Y, Huang X, Zhang Y & Wang S 2008 Phay JE & Ringel MD 2013 Metastatic mechanisms in follicular cell-derived Overexpression of macrophage migration inhibitory factor induces thyroid cancer. Endocrine-Related Cancer 20 R307–R319. (doi:10.1530/ angiogenesis in human breast cancer. Cancer Letters 261 147–157. ERC-13-0187) (doi:10.1016/j.canlet.2007.11.028) Qi D, Atsina K, Qu L, Hu X, Wu X, Xu B, Piecychna M, Leng L, Fingerle- Zheng YX, Yang M, Rong TT, Yuan XL, Ma YH, Wang ZH, Shen LS & Cui L Rowson G, Zhang J et al. 2014 The vestigial enzyme D-dopachrome 2012 CD74 and macrophage migration inhibitory factor as therapeutic tautomerase protects the heart against ischemic injury. Journal of targets in gastric cancer. World Journal of Gastroenterology 18 2253–2261. Clinical Investigation 124 3540–3550. (doi:10.1172/JCI73061) (doi:10.3748/wjg.v18.i18.2253)

Endocrine-Related Cancer Received in ﬁnal form 5 January 2015 Accepted 19 January 2015 Made available online as an Accepted Preprint 19 January 2015